SuperedgeRank algorithm and its application in identifying opinion leader of online public opinion supernetwork

Opinion leaders on the internet are very important figures in online communities, which play an important role in promoting the formation of public opinions. Many theories have been introduced to identify opinion leaders by social network analysis, text mining and PageRank-based algorithm in different fields, but few has addressed the issue of opinion leader identification by combining the methods above, and there is no research using supernetwork analysis to identify opinion leaders. This paper proposed an SuperedgeRank algorithm for opinion leader identification based on supernetwork theory, which combined the network topology analysis and text mining. First, the study established a supernetwork model with multidimensional subnetworks, which are social, psychological, environmental and viewpoint subnetworks. Then, the study proposed four supernetwork indexes: node superdegree, superedge degree, superedge–superedge distance and superedge overlap. The later two indexes are developed by us to help evaluate the identified opinion leaders. Based on them, our study applied SuperedgeRank algorithm to rank superedges, and used the ranking result to identify opinion leaders in opinion supernetwork model. Finally, the feasibility and innovativeness of this method were verified by a case study.